April 20, 1955
PARTIAL
[CONTRIBUTION FROM
THE
HYDROLYSIS OF GERMANIUM TETRACHLORIDE
DEPARTMENT OF CHEMISTRY, MASSACHUSETTS INSTITUTE O F
2 1:33
TECHNOLOGY]
The Partial Hydrolysis of Germanium Tetrachloride BY WALTERC. SCHUMB AND DONALD M. SMYTH RECEIVED OCTOBER15, 1954 The partial hydrolysis of germanium tetrachloride at -78” in various organic solvents has led to the isolation of only solid, polymerized products with compositions close t o GezO&ln. The hydrolysis reaction was found to cease when the aqueous phase had absorbed a sufficient quantity of the liberated HCl to bring its concentration t o about 6.5 molar. Calculations based on an analytical examination of the reaction mixtures indicate that small amounts of one or more members of what would be a homologous series of oxychlorides, Ge,O,- 1C1:,+ 2 , are also possible products of the hydrolysis.
Recent work in this Laboratory has served to arouse considerable interest in the oxyhalides of silicon, which are compounds taking the form of a homologous, linear series, SinOn-lX~n+2, and a cyclic tetramer, (SiOX2)4,where X represents chlorine or bromine and n is an integer greater than one.2--j Of particular interest is the method of preparation conceived by Schumb and Stevens,j which involves the controlled, low temperature hydrolysis of silicon tetrachloride in ethereal so1ution.j This technique has been further extended by Schumb and Lefever in their partial hydrolysis of hexachlorodisilane, whereby a new series of silicon oxychlorides with the general formula Sizn+20nC14n+6 has been shown to exist.6 I t is interesting to note that cyclic oxychlorides have not been isolated from either of these reactioW.”-’ In a recent publication Schwarz and K n a d have described the partial hydrolysis of tetramethoxygermane and tetraisopropoxygermane.8 Hexaisopropoxydigermoxane, Ge20(OC3H7)6, was isolated from the latter reaction by very careful vacuum fractionation. Polymerized solid products with compositions approaching Ge203(OC3H,)2were also reported. Two oxychlorides of germanium have been reported in the literature. Hexachlorodigermoxane, Ge20Cls,has been prepared by the lengthy reaction of GeC14 with oxygen a t ~ F I O O . ~ This preparation has been repeated and confirmed in the course of the present work. Another oxychloride, GeOC12, has been reported as the product of the reaction of trichlorogermane with silver oxide. lo This reaction was also repeated and the authors have submitted a note for publication which indicates that the earlier work was in error. The reported existence of GezOC16, which would be the first member of a homologous series of germanium oxychlorides, GenOn-1Clan+2, suggested that the partial hydrolysis of GeC14 might yield this compound as well as higher members of this (1) Based on part of a thesis presented by D . M. Smyth t o the Department of Chemistry, Massachusetts Institute of Technology, in partial fulfillment of the requirements for the degree of Doctor of Philosophy. (2) W. C . Schumb and C. H. Klein, THIS JOURNAL, 69, 261 (1937). (3) W.C.Schumb and D. F. Holloway, ibid , 63, 2753 (1941). (4) W. C. Schumb, Chem. Revs., 31, 587 (1942). ( 5 ) W. C. Schumb and A. J . Stevens, THIS JOURNAL, 78, 3178 (1950). (6) W. C.Schumb and R. A. Lefever, ibid., 76, 1613 (1953). (7) J. Goubeau and R. Warncke, 2. anorg. Ckem., 269, 109 (1949). (8) R.Schwarz and R. G. Knauff, 2. anorg. aiigem. Ckem., 271, 193 (1954). (9) R.Schwarz, P. W. Schenk and H. Giese, Ber., 6 0 , 362 (1931). (10) R. Schwarz and F. Heinrich, Z. anorg. allgem. Ckem., 209, 273 (1930).
series. On the contrary, it was found that no volatile oxychlorides of germanium could be isolated from this reaction, although small quantities seem to have been formed. The principal product was found to be a polymerized solid with the composition Gez03C12.
Experimental The partial hydrolysis technique has been thoroughly described in earlier p~blications.6,~As applied here it was carried out a t -78’ in one of the three solvents: ethyl ether, chloroform or a-pentane. The amount of water employed was varied between GeCI,: HzO molar ratios of 5 : 1 and 1 : 1 with values near 2 : 1 being more commonly used. The variation of solvent: GeCI4 ratio was found to have no observable effect down to volume ratios of 2 : 1 which was the minimum used. The GeCI4 was prepared both by :he reaction of Ge02 with hydrochloric acid as described in Inorganic Syntheses,”’l and by the direct chlorination of the metal. The latter is the more efficient method if the metal is available. The product was carefully fractionated from anhydrous NazCO, to remove HC1 when the first method was employed, and from HgZCI? to remove CL when the direct chlorination was used. The samples were analyzed by t h e following method. Samples were weighed in small, thin-walled glass bulbs which were then broken in a n excess of water in a stoppered flask. Chloride was determined by titratio.1 with standard base, after which the solution was boiled to dissolve GeOt and 10 g. of d-mannitol was added. When cool, the germanium dioxide-mannitol complex acid was also titrated with standard base. A Beckman Model G pH meter was used t o determine both end-points. The oxygen contents of the products were calculated from the valence requirements. S o evidence was ever noted of degradation t o lower valent germanium or to the metal. Molecular weights were determined cryoscopically in benzene. The solvents were Mallinckrodt analytical reagent grade ethyl ether, chloroform and benzene, and Phillips pure pentane. All of these were found to be satisfactory without further treatment. Lengthy manipulations of the hydrolyzable materials were carried out in a carefully maintained dry-box. It was found, however, t h a t many steps could be carried out in the open air without introducing detectable errors if the exposure was kept at a minimum. The fractional distillations were carried out in a 50 X 10 cm. column packed with S/,2-inch Fenske helices.
Results and Discussion Repeated attempts to, isolate volatile germanium oxychlorides from the partial hydrolysis products by low pressure fractional distillation were unsuccessful. I n every low pressure fractionation of the hydrolysis products only unreacted GeCt and the solvent were distillable, and a large non-volatile residue remained which consisted of an extremely viscous oil or glassy solid. Removal of the volatile materials by suction a t room temperature gave the same results. When the volatile materials from (11) W.C.Fernelius, “Inorgsnic Syntheses,” Vol. 11, McGraw-Hill Book Co., Inc.. New York, N.Y . , 1946, p . 109.
four successive hydrolysis reactions were combined tion. No such appearance was noted in (;c?OaCl2 and fractionated, a sinal1 quantity of non-volatile and this may be due to a more random type of polysolid was ol)tainc.tl, indicating that some volatile merized structure. n'ork in this Lahorattrry on the species was formcd wliich was decomposed by the thiohydrolysis of Sic14 has led to thc isolation nf heat of the suhsecltir-nt distillation. This volatile solid products having compositions very close to material could have been (k20C16which has been Si2S3C12.'2The so-called silicoforniic anhydride, fourid to be thernially unstable to a coiisiderahle ex- Si203H2, is also of similar composition. These extent. Ii'hen prepared by Schwarz's method,Y amples indicate that this composition has 21 niuch this compound was isolated in only small quantities greater generality than has been previously realized. by distillation from ten to twenty times as much I t was found in the hydrolyses in which the solcrude product. Deconiposition was also indicated vent was pentane or chloroform, that much of the 1)y the tliffculty in obtaining a consistent boiling water was not consumed by the hydrolysis reaction. point for the CJe2OCls. The best value seemed to I n ether solution a much greater quantity of water be .i(joat 1:1.5 mni., which is considerably higher was consumed by the reaction. This phenoinenon than the previously reported value. On redistilla- was attributed to the fact that GeC14is not hydrution a t 13.3 nim. a ten-gram s-tmple of GezOC16 de- lyzed by relatively concentrated solutions of hydrocomposed completely to a glassy white solid. This chloric acid.l33l4 This is a marked difference from behavior was strikingly similar to that of the hy- the case of SiC14. Apparently the HC1 liberated by drolysis products and is in accord with the concept the hydrolysis was rapidly dissolved into the aquethat GenOClr, may have been a minor hydrolysis ous phase until its concentration reached the limitproduct. ing value, at which point the hydrnlysis ceased. The non-volatile residues varied considerably in An examination of these unreacted aqueous phases consistency, and analyses showed a parallel trend showed this equilibrium concentration of hvdrrlwhereby the more viscous materials were relatively chloric acid with CIeCl, to be (i.,)O..) inolar in the p o r e r in chlorine content. I t was found that ex- reactions in which ethyl ether or chlornform were traction with anhydrous benzene left a white, in- the solvents. Determinations in pentane solution soluble solid, arid the analyses of these residues were handicapped by the fact that emulsions were showed compositions very close to Ge203C12,depend- formed of the pentane phase in the aqueous phase. ing on the eficiency of the extraction. A s ex- Presumably the extent of hydrolysis could be inamples, one product which was extracted for one creased by the use of larger amounts of water or by day at room temperature gave a residue with a the addition of an acceptor for the HC1. The C1: Ge atomic ratio of 1.10 : 1 .OO; another which was limiting concentration of HC1 at which hydrolysis extracted for a day with hot benzene in a Soxhlet of GeCll ceases is in reasonable agreement with extractor gave a ratio of O.B7:1.00. These corre- other published work in which the limiting value spond to approximate formulas of Gezo0O2.9oC1~.?o was found by less clircct means. 1:1,11 arid Ge2.0~Os.n:~Cll !v~ respectively. , The change of Ethyl ether is a much better solvent for HC1 than composition due to the extraction indicated that chloroforni and Iz-pentane, :mtl for this reason the relatively chlorine-rich inaterial was being removed hydrolysis procceded to a inuch greater extent in (presumably GeC14and possibly Ge20Clnor higher this solvent, since more HC1 was required to make ineinbcrs of this series). A number of these experi- the aqueous phase 6.,5 molar in HC1 in equilibrium ments gave similar results regardless of whether with the ether phase. The water was completely ethyl ether, chlor!)form or wpentane was employed consumed by the reaction in ether solution when as the solvent during the hydrolysis. The solid dis- the initial GeC14:H 2 0 molar ratio was greater than tillation residue from a preparation of Ge2OCle about 2 . 2 : 1.0. On the other hand, the reaction by the thermal oxidatim of GeC14 was also sub- was not complete in the other two solvents even jected to this extraction treatment and yielded a a t GeC14:HzOratios as high as 5.0 : 1.O. similar residue with the composition Gez,oo0308I t is interesting to note that in the case of all Cll.84. This is in agreement with the earlier experi- three solvents used in this study, ethyl ether, ?I.ments which indicate that Ge20C16 is transformed pentane and chloroform, only minute traces of into the solid polymer by the action of heat. GeOz were observed as a hydrolysis product. In 'I'liese solids were found to retain the benzene the partial hydrolysis of sic14 it has been shown very tenaciously, and even several hours of pump- that ethers are the only suitable solvents, since a i:I,q would not remove the last few percentage of wide variety of other solvents led to the formation solvent. For this reason the compositions were of only ~ i l i c a . ~ An analytical study of this reaction has further derived from the C1:Ge ratios rather than the absolute percentages. This was justified by the relia- clarified the nature of the hydrolysis products. For bility of the analytical method as indicated by its this purpose several hydrolyses were carried out, wcuracy when used on samples not containing and after the solvent was removed by distillation, occasional samples were withdrawn from the distillhenzene. The general composition of GenOaClzfinds some ing pot while the unreacted GeC14 was being reinteresting parallels in other work. Schwarz has moved. These samples were analyzed for chloreported that the partial hydrolysis of tetraisopro- rine and germanium and the molecular weights were determined by the freezing point depression of benpoxygermane leads to considerable quantities of a W. J. Bernard, P h . D Thesis, hI.I.T., June, 1 R . X solid polymer with the composition GezOB(OC3H7)2.* (12) (13) E. R. Allison and J. H. Milller, THISJ O U R N A L , 84, 2838 (1932). Schwarz describes this product as having a fibrous (14) G H. Xforrison, E G. Dortman and J . F. Cusgrove, ;bid., 76, structure and postulates a linear type of polymeriza- 4236 (1934).
*
P.ZRTI.IT, IIYDROLYSIS 01;GERMANIUM TETR.ICIII.ORIDE TABLE I Sample
CI
Ge
A1 A2 B1 B2
1 ,797
11.182
1.762 1 ,828 1.711 1.627 1,690 1,377 1 ,642 1.613 1.571
,490
1. 5Ofl
. ,569
c1 c2 C8 I) 1 D3 I ):1 IN
Total % accounted for
99.74 9 9 , 6.3 1no.71 90 . 9 1
,184 , XI5 ,530 ,534 ,541 ,528 , 5.32
loo, os 99.98 99.99 99.8’;
ioo ,no
. ,544
100.20 100.76
--n
M
239 2.50 280 265 31% 339 364 29;i 311 336 388
zene. The oxygen content could be calculated by the valence requirements and the percentage totals thus accounted for are shown in Table I. If the products are assumed to be composed of only (Ge203C12)a and unreacted GeC14, the concentration of the oxychloride and its degree of polymerization a can be calculated from the data. I n each case negative quantities of (Ge20jC12),, and negative values of a are required to satisfy the data. The postulate is obviously incorrect. If, on the other hand, the products are assumed to be (Gez03Clz)aand members of the linear homologous series, GenOn--1C12n+2 (which includes GeCL as the case n = 1)) the relative amounts of each can be calculated along with corresponding values of n for various values of a,the degree of polymerization of the (GezO&lz)a. The results of these calculations are shown in Table I. A and B represent two hydrolyses in which n-pentane was the solvent, and C and D were two reactions in which ethyl ether was the solvent. Only two samples were withdrawn from each of the pentane solutions because of the relatively smaller extent of hydrolysis in this solvent as explained earlier. The calculations were based on the following formulas, which can be easily derived. n =
+
-
2a[200 M(Ge Cl,] - 2AfGe Af(2Ge - C1) - 200
x =
Ge(n
+ 2) - n(C1 - Ge) 2a(n + 2)
a = 4
a - 1
(1)
(3 )
In these equations y represents the concentration of Ge,rO-n-1C12-n+z in moles per 100 g., and x denotes the concentration of (GezOaClz).. C1 and Ge represent the gram atoms per 100 g. of the two elements and ,I1 is the experimental molecular weight. a and n have the same significance as in the preceding discussion. a has been left as an adjustable parameter in the equations. I t is easily seen that the value of a has very little effect on the calculated quantities, except, of course, the concentration of (Gez03Clz)a,which is inversely proportional to a. This means that the calculations offer no useful information as to the degree of polymerization of the Ge2O3Cl2. All of the derived values and the trends shown in each experiment (with the single exception of the value of x for sample C-1) are consistent with expectations. Thus n is always slightly greater than one and increases as GeCL is removed, and the concentration
1 .ltj 1.21 1.11 1.31 1.64
1.1.5 1.19 1 .00 1.31
1.80 1.97 1.51 1.65 1.81 2.21
1.69 1.85 1.3s 1.49 1.58 1.80
l.jl
i i = m
1.13 1.18 1 .os 1.20 1.43 1.62 1.77 1.30 1 ,-lo 1,45 1.57
213.5
Concn. of product.s, moIes/100 g. (a = 1) (GezOrCId Ce.0- iChn+ z
0 0008 .nn54 ,0028 ,0132 ,0179 .0141
,0142 0214 ,0219 , ,282 , 0305
n.418 ,396 ,430 ,363 ,302 ,281 ,261 ,319 .296 ,270 ,221
of GenOn--1C12fl+2 decreases and that of (Ge203C&), increases as GeCh is removed. Thus the postulated product mixture is compatible with the data. The indication is that small quantities of the linear oxychlorides are formed, and this is in agreement with the experimental fact that some volatile products were detected which polymerized during distillation attempts. The bulk product appears to be the material (Ge203C12),with a having possible values from one to infinity (presumably as a approaches infinity the material would become colloidal in nature), The observation that the viscous, non-volatile products had a pronounced tendency to form gel-like structures with benzene lends some support to the idea that the product is present as a colloidally dispersed phase. Differential error functions have been derived for equations 1, 2 and 3, and their evaluations a t the experimental points with assumptions of reasonable error limits have indicated that the calculated values are significant. The authors wish to point out very clearly that the assumptions made above are not unique in fitting the data. Several hypothetical mixtures can be proposed which are equally compatible, but the products discussed above are much more obvious choices in view of the experimental facts and the chemistry of analogous silicon compounds which can often profitably be compared with that of germanium. I t should also be stated that the indicated amounts of the linear compounds, Ge-nO,-lClZfl+2,could very well be expressed entirely in terms of the known member of this series, GezOCle, but the more general case has been used here. The germanium analog of trichlorosilanol, SiC1,OH, has not been observed during this investigation. The silicon compound was reported by Goubeau and LVarncke’ who isolated it from the partial hydrolysis of Sic14 by fractional distillation a t atmospheric pressure, and by EmelCus and RTelchl5from the photooxidation of trichlorosilane. Actually, serious discrepancies exist between the descriptions of this compound provided by the two separate investigations. The presence of the germanium analog, once the hydrolysis solvent has been removed and the product has reached the boiling point of GeCL, was eliminated by the analytical data which indicate that the product consists only of germanium, chlorine and oxygen a t this point. An additional study has shown that the water which is consumed is essentially quantitatively con(15)
H.J. EmelCus and A. J. E.Welch, J . Chcm. SOC.,1928 (1939).
verted to HC1 by this point in the experiment. -1 weighed amount of water was used for a hydrolysis and 98.6‘;; of its theoretical equivalent as HC1 was captured and determined during the distillation. Both of these facts show that there is little hope of isolating GeClnOHfrom this reaction by distillation, but they do not eliminate its existence tluriiig the
[CONTRIBUTIOS FROM THE
DEPARTMENT OF
earlier stages of the reaction where it may be formed and then consumed by rapid condensation. Acknowledgment.-The authors express their appreciation to the Lincoln Laboratories a t this Institute for their interest and material aid in this study. CAMBRIDGII,
h~.iSS.iCHL’SETTS
CHEMISTRY, U S I V E R S I T Y O F I ’ E N S S Y L V A S I A ]
Some Heteropoly 6-Molybdate Anions : Their Formulas, Strengths of their Free Acids, and Structural Considerations1 BY LOUISC. IT.BAKER,’GEORGE FOSTER, WILLYTAN,FRANK SCHOLKICK A N D Tiioms P. ~ ~ ~ C U T ~ I ~ E O S RECEIVED OCTOBER 21, 1954
A study of 6-molybdo heteropoly anions, in which tervalent aluminum, chromium, iron or cobalt functions as central atom, indicates that [(XO&fo60~,)~] --3n is a reasonable formula for these anions, rather than [X+S(Mo04)sj -’3, as given in an older notation. XO6 is an octahedral group containing the central atom and n is an undetermined integer which is probably small but probably not unity. Structural considerations, which are included, suggest that the compounds may be bimolecular. A method is given for the preparation of ammonium 6-molybdoferrate(III) in a pure state. The free acid of the aluminum complex is described. I t is made evident that there is neither a general formula nor a single structural model which can represent 6-heteropoly anions as a class.
Recent work” on the determination of the valence of a heteropoly anion of the 12-molybdo class suggested a restudy of 6-molybdo heteropoly anions containing tervalent aluminum, chromium, iron and cobalt as central atoms. These ii-molybdo compounds, discovered a century ago,* were the subject of several early inve~tigations,j-~ in whch dualistic “oxide” formulas were used. Miolati-Rosenheima-’’ formulas, modifications thereof, and formulas implying the same anionic composition have been widely used by recent to represent these and other 6-heteropoly compounds. Examples of such formulas are: (SHq)8H~[X(hloOl)s].10H20 or (NH4j3H6[XMo6024]~10H20, where X represents one of the tervalent elements listed above. In 1929, Pauling la accepted Rosenheim’sformulas (1) Presented before t h e Physical and Inorganic Divisiun of t h e hleeting-in-hliniature of t h e Philadelphia Section. American Chemical Society, January 18, 1931 (3) Department of Chemistry, Boston Vniversity, Buston 1.5, Mass (3) L C W. Baker, 0 A. GallaRher and T. P. XIcCutcheon, ’THIS JOURNAL, 76, 2493 (19.531 (4) H . Struve, J . p r u k f Chem., 61, 44!) (18541 ( 5 ) J . G. Gentele. i b i d , 81, 413 (1860). (6) E 1Iarckwald. D i s e r t a t i o n , University of Rase!. Switzerland 1893 (7) R . D. Hall, Trrrs J O U H N A L , 29, 692 ff. (1907). (8) A. AIiolati, J 9 , a k t . C h e m . , 77, 417 (1908). (9) A. Rosenheim and H . Schwer. Z. u , z o v ~ .C h e v z . , 89, 224 (1914). (10) A Rosenheim in A b e y g ’ s “Handbuch der anorg. Chem .” Val, IV, part 1 . i i , Leipzia, Germany, 1921. pi> 977-106.5 ( 1 l i (a) J. W hlellor, “.4 Comprehensive Treatise un Inorganic and ’Theoretical Chemistry,” VIA X I , Inngmans, Green and eo., lmndun, 1931, pp, 600-603. ( h ) Gmelins “Handbirch der anorc Chem ,” 8th Edition, Vol. .73, 374 (1933). (12) A . F. Wells, “Structural Inorganic Chemistry,” Oxford University Press, New York, N. Y . , 1947, pp. 344-34.5. (13) H . J . Emelkus and J S. Anderson, “Modern Aspects of Inorganic Chemistry,” Secnnrl Edition, D . Van Nostrand Co., Inc., New York, K. Y . , 1982, pp. 22.5-226. See also First Edition, p p . 193-194 (14) H . T . Hall and H . Eyring, THISJ O U R N A L , 72, 782 (19.70). (15) I . Lindqvist, Arkia K c m i , 2, 325 (1950). (16) P. RAY,A. Bhaduri and B . Sarma, J . Indinri C h e m Soc.. 2 6 , .jl (1948). ( I 7) I,. C. W. Baker, B Loev and T. P. IvlcCutcheon, THISJOUHKAL, 72, 2374 (1950). (18) L . Pauling, ibid., 51, 2868 (1929).
and structure for the 6-heteropoly anions considered herein. In the same paper he rejected the XiolatiKosenheim formulas for 12-heteropoly compounds and made the fundamental suggestions concerning the considerations governing their structures which, after certain important modifications, led to the “Keggin Type” formula for 12-heteropoly anions.3,19-?3 In 1937, Anderson13,24 proposed a structure for 6-heteropoly anions based on the principles made evident by the above-mentioned studies of 12heteropoly compounds. The central atom in the .Anderson model is in octahedral coordination with oxygen, and the central octahedron is surrounded by a hexagon of Moo6 octahedra. All of the octahedra lie in one plane. The anion, as in the Kosenheim formula, contains twenty-four oxygen atoms. This hypothetical configuration received experimental support when EvansZ5 showed by X-ray analysis that it is an extremely probable structure for the 6-molybdotellurate(V1) anion in (NH4l6[Te06M060~8].7Hz0 and & [Te0&\1060,~]. TH2O. Evans has recently reported further X-ray evidence which conclusively establishes his former The work described in this paper, on 6-molybdo complexes of tervalent metals, is not in agreement with the views of Rosenheim and S ~ h w e r . Their ~ conclusions were based on conductivity measurements (which Rosenheim observed might also be interpreted as consistent with normal salt formulas 1 and, especially, on dehydration data. Their results (19) J. F. Keggin, Satrrrc, 131. 908 (1!l3R) (20) J. F. Keggin, Pvoc. R o y S C J C ( L m ~ d o t i )A144, , 75 (19343. (21) (a) J. Id. Hoard, Z.Krisl , 84, 217 (1933): (h) J. A. Santos, PYOC.R o y . Soc. ( L o n d m ) , 8160, 30Q (19351: ( c i 0 . Kraus, Z . Krisl 91. 402 (1935); B8, 379 (19361. (22) R. Signer and H. Gross, Hels Chim. Acta, 17, 1076 (1934). (23) J. W. Illingworth and J . F . Keggin, J . Chem. Soc., 576 (193.5). (24) J. S . Anderson, Srilrrre, 140, 830 (1937). (25) H. Evans, J r . , T H r s J O U H S A L , 7 0 , 1291 (1948) (26) H. Evans, J r , ahstractr uf papers presented before the spring meeting of the American i‘rystall
